Open Access
Add to BibSonomy
Abstract
As the use of photonics circuits expands, the optical quality and performance of integrated components in the microscale become a major concern. Aiming to improve the performance while reducing the time processing, new microfabrication approaches are being investigated. The dewetting of glassy thin films have been recently proposed as an alternative for nano and microfabrication of chalcogenide optical components. Besides being the best materials for light transmission in the infrared region, chalcogenide glasses possess a flexible molecular structure that allows using a cheap and simple molding process. Here we investigate the thermal-induced dewetting of chalcogenide As20Se80 thin films, by studying the influence of temperature, atmosphere, and heating rate on the formation of self-assembled microstructures. We found that thin films between 150 and 700 nm dewet via structural relaxation, similarly to liquid agglomeration, and produce solid microstructures with the same composition and molecular structure as the initial film. By controlling the glass viscosity and the kinetics of the nucleation process it was possible to adjust the distribution and size of glassy microstructures. Additionally, we combine the dewetting process with standard photolithography and by avoiding the capillary instabilities, we are capable to obtain waveguides with the smooth and symmetric surfaces required for optical applications in the microscale size.
© 2021 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
The chalcogenide glasses are the best materials available for the transmission and manipulation of infrared light. Their unique properties of transparency in the mid-infrared region, high refractive index, and high non-linearity have been highly exploited for telecommunications, optical sensors, ultrafast not linear optics, among others [1–3]. In the field of integrated photonics, they have been proven as low loss waveguides, microresonators, and active media in on-chip photonic devices [3–6]. However, their use in integrated circuits remains a challenge due to the complexity of miniaturization and the integration with other optical components. To fabricate chalcogenide microstructures through conventional photolithography is unsatisfactory and difficult. Besides the struggles of performing uniform dry etching, these materials are chemically attacked by the alkaline resist developers used in wet etching. These processes result in asymmetric structures with high side roughness and with dissimilar sizes that cause high optical losses.
The dewetting of thin films is a promising alternative for the creation of micro and nanostructures of chalcogenide in solid substrates. Driven by the minimization of surface energy, the homogeneous layer of material can spontaneously break to form separated islands or droplets of material which dimensions are in the micro and nanoscale. This phenomenon is governed by the interface interactions that take place between substrate, film, and atmosphere, and by the molecular interactions within the film material [7,8]. Due to the molecular structure of chalcogenide glasses, they are the only family of inorganic glasses that are known to dewet at low temperatures, while maintaining their molecular structure and consequently their optical properties. Furthermore, the auto-assembled microstructures of glass produced during dewetting have always smooth and symmetric surfaces suitable for use in optical applications.
Although exists substantial evidence that sets the process of dewetting as an important tool for micro and nanofabrication of metals and polymers [9–11], its use in photonics has been little explored. As a proof of concept, the dewetting of chalcogenide glasses has been recently used to create photonic structures of high optical quality in dielectric metasurfaces and etchless microresonators [5,12]. However, to take full advantage of dewetting to produce photonic structures with desired sizes and specific spatial distribution, it is important to understand the influence of the film's molecular structure, atmosphere, and thermal treatment.
In this work, we explore the influence of the thermal treatment conditions and film thickness in the dewetting of As20Se80 films. These Selenide-rich glasses have a high refractive index and a wide transmission window, which make them good candidates for waveguides and active components in optical circuits [2]. Additionally, our previous work showed that dewetting of As20Se80 films is expected to occur at temperatures lower than 300°C, without crystallization, and with negligible changes to their molecular structure [13]. To overcome the limitations of photolithography and produce structures more suitable for optical applications, we thermally induce the dewetting of microstructures formed by photolithography. By engineering the mass transport mechanism together with the capillary instabilities during dewetting, it is possible to create highly symmetric structures with varied shapes and sizes in the microscale.
2. Experimental
As20Se80 thin films were deposited on clean microscope slides (Fisherbrand) through electron beam evaporation, in a Nanochrome II electron-beam IntlVac system under high vacuum conditions (10−6 Torr). With the help of a quartz crystal sensor, the deposition rate is monitored and maintained at 10 $\dot{A}$/s until thickness between 150 and 2500 nm are achieved. The initial As20Se80 bulk glass for the evaporation was obtained by the melt-quenching method, in which high purity arsenic and selenium are sealed in a silica ampoule under vacuum and then heated for 14 h at 850°C.
The main part of the experiments were performed using microscope slides as substrate. The slides are made of soda-lime glass which specific composition was verified by Energy-dispersive X-ray spectroscopy (EDX) and reported in the Supplement 1. To examine the influence of the substrate composition and the process of dewetting in substrates closer to application, films deposited in quartz, silicon, and thermally grown silica are also studied. All the substrates were cleaned in an ultrasound bath with water, acetone, and isopropanol. Values of roughness were calculated from Atomic Force Microscopy (AFM) images acquired in tapping mode using an AFM Dimension V SPM model from Veeco.
To induce the dewetting, the films were placed inside of a heating stage, in which the temperature, heating rate, and atmosphere can be controlled. The stage is coupled to an optical microscope that allows monitoring the surface of the film during the process of dewetting. To extract information on the film agglomeration kinetics and final state of droplets distribution, some images are acquired along the process and then analyzed with ImageJ software. The dewetted area fraction is defined by $X = {A_H}/{A_T}$ (with ${A_H}$ the dewetted area in a reference region of area ${A_T}$), while the mean size of the droplets refers to the mean ferret diameter. Moreover, the correlation in the position of the droplets is analyzed by extracting the position of the droplets and performing the 2D Fourier transform of the spatial distribution.
The composition of the “as-deposited” films and the droplets of dewetting were analyzed in a FEI Quanta 3D FEG system coupled with an Energy-dispersive X-ray spectroscopy (EDX) detector, using between 15 and 20 KV energy. The same microscope was used to obtain the SEM images of the As20Se80 microstructures. The changes in the glass molecular structure due to dewetting were examined by Raman spectroscopy in a Renishaw inVia equipment coupled with a Leica DM2700 microscope and a 785 nm laser as the excitation source.
We used X-Ray photoelectron spectroscopy (XPS) to understand the influence of the atmosphere on the dewetting process. Three films were placed inside of the XPS equipment immediately after deposition. One of the films is analyzed directly as-deposited, while the other two were treated inside a high pressure-temperature reactor coupled to the XPS analysis chamber. To reproduce the interactions that can take place during dewetting, the two samples were treated at 100 °C for 30 min under 760 Torr of nitrogen or oxygen, respectively. The surface of the films was analyzed in a Kratos axis ultra XPS system equipped with a monochromatic x-ray source Al Kα (1486.6 eV). Corrections due to surface charge effects were performed using the C 1s peak of adventitious carbon (284.6 eV) as a reference and the background subtraction in the high-resolution core levels is made through the Shirley method. The fitting of As 3d and Se 3d spectra is performed maintaining the area ratio between the spin-orbit splitting components 3d3/2 and 3d5/2 of As and Se.
3. Results
3.1 Dewetting of As20Se80 thin films
The dewetting of glassy films with thicknesses between 150 and 700 nm is achieved by heating the samples at temperatures between 180 and 300 °C (Tg+79°C and Tg+199 °C respectively) in a nitrogen atmosphere. The time for the total dewetting decreases with the temperature due to its influence on the viscosity of the glass and hence on the velocity of material agglomeration.
Regardless of the temperature, substrate and film thickness, the process by which the film breaks and agglomerates seems to be the same. As the temperature increases, some substrate-exposed regions appear at randomly distributed positions of the film surface. The area of such holes grows as the film’s material retract and accumulates in its edges (rims). When bumping with other holes, the accumulated material forms a ribbon that eventually decays into droplets due to capillary instabilities. The collision of circular holes induces a distribution of droplets in the polygons-like pattern observed in Fig. 1, which is similar to the patterns usually observed in polymeric thin films [14,15]. At the initial stage of dewetting, the holes occupy positions that are not spatially correlated, this indicates a rupture of the films (with thicknesses between 150 and 700 nm) by the heterogeneous nucleation mechanism.
Fig. 1. Optical images of As20Se80 dewetted films when submitted at 300 °C. The film thicknesses are 150 (a), 225 (b) 340 (c), and 510 nm (d). The scale bar is 500 µm.
Regarding the composition and molecular structure of the droplets of glass after cooling, there are not significant differences compared to the film's initial material. While the EDX measurements showed no significant variations in the film's atomic composition, the Raman spectra of the islands showed a band identical (in shape and position) to the “as-deposited” film. Further analysis allows us to deconvolute the Raman spectrum in the three different peaks shown in Fig. 2. The peak around 224.1 cm-1 corresponds to vibrations of the pyramidal structural units AsSe3, where the arsenic atom is at the top of the pyramid connected to three Se atoms. The other two peaks, around 275.0 and 253.7 cm-1, are commonly observed in Selenium rich-glasses where the excess of selenium (out of stoichiometry As2Se3) is forming selenium chains that connect the pyramidal units AsSe3 or forming domains of amorphous selenium in Se8 rings respectively [16,17].
Fig. 2. Raman spectra of dewetted material obtained by heating As20Se80 films at 300 °C. The experimental Raman band is deconvoluted by using three main peaks.
Compared with the dewetting on the soda-lime glass substrate (microscope slides), chemically similar substrates (quartz and thermally grown silica on silicon) show little to no variation in the diameter and the contact angle of the droplets (see figure S1 and S2 in Supplement 1). All tested substrates presented a roughness smaller than 1.5 nm (Sq) and the rugosities of their respective lie below 1 nm. Such does not influence the final dewetted droplets. We concluded that, despite their structural and chemical differences, the dewetting is produced similarly for all three substrates. For simplicity, the following experiments were performed in films deposited on soda-lime microscope slides.
Although dewetting at different temperatures and thicknesses does not cause chemical and structural changes in the film material, these parameters are responsible for extensive variations in the mean size of dewetted islands and the dynamics of the process. As shown in Fig. 3(a), the characteristic temperature at which the first holes of dewetting appear (Tdewet) is thickness-dependent. It decreases for films thinner than 340 nm and maintains a value around 215 °C for thicker films. The tendency (dashed line of Fig. 3(a)) has also been observed in several metallic and polymeric systems in which the interfacial interactions with the substrate become more relevant as the thickness decreases. This causes values as the viscosity, glass transition temperature, and activation energy for viscous flow to be thickness-dependent [18,19].
Fig. 3. Temperature at which the first holes of dewetting appear upon heating at 10 °C/min (a). Mean drop diameter and number of holes when inducing dewetting at 250°C in films with different thicknesses (b). The dashed lines are only to guide the reader’s eyes.
The dewetting of As20Se80 films also produces droplets of material whose size varies linearly with the film thickness (Fig. 3(b)). By inducing dewetting on films with thicknesses between 150 and 700 nm, structures with diameters ranging between 13 and 84 µm are formed. The size of the droplets is also inversely proportional to the number of holes that appear in the early stages of dewetting (Fig. 3(b)). The density of holes determines the amount of material that accumulates on the border of each hole before colliding with others and form the droplets of dewetting. Although the size of droplets also depends on the conditions of the temperature treatment (section 2), the coarse tuning of droplets’ size is indeed achieved by changing the film thickness [8,11,18,20].
3.2 Isothermals of dewetting
To understand the influence of the temperature in the process of dewetting, we submitted the samples at constant temperatures, between 200 and 300 °C until the total dewetting of the films. The samples are heated rapidly at a constant rate of 100 °C/min and at the isothermal temperature they are monitored optically. As expected, the kinetics of the process and the final size of islands are greatly influenced by the temperature since the viscosity of the glassy film is temperature-dependent. High temperatures imply lower viscosities which ease the agglomeration of material and allow the formation of bigger drops. As observed in Fig. 4, the size of particles rises with the isothermal temperature at which were formed. By inducing dewetting in films of 510 nm, it is possible to obtain islands of material with mean sizes ranging between 64 µm and 80 µm. Due to the natural enhancement of viscosity in thinner films mentioned before, the influence of temperature has slight effects on the thinner films and becomes more important at thicker ones (see Fig. 4).
The temperature also affects the kinetics of the dewetting registered in the transformation curves of Fig. 5(a). Here the evolution of dewetted area fraction is recorded for different isothermal treatments. Considering that dewetting occurs in a free isotropic surface due to spontaneous rupture of the film at randomly distributed sites and that holes grow constantly on time, it is possible to fit the experimental data of Fig. 5(a) by using the equation [21]:
where X is dewetted area fraction, $\tau $ is the delay time for nucleation and K and R are defined from the number of holes per unit area (${N_H}$), the velocity of dewetting ($u$) and the initial radius of the holes (${r_0}$) as follows: $K = \sqrt {\pi {N_H}} u$ and $R = \sqrt {\pi {N_H}} {r_0}$. This model, successfully used in the dewetting of As-Se glasses [13], fits our experimental data in the dewetting of 150 nm film (dashed lines in Fig. 5(a)). Knowing that the velocity of dewetting u depends on the temperature T and the activation energy for dewetting (${E_a}$) as $u \propto \frac{{exp({{\raise0.7ex\hbox{${ - {E_a}}$} \!\mathord{\left/ {\vphantom {{ - {E_a}} {{k_B}T}}} \right.}\!\lower0.7ex\hbox{${{k_B}T}$}}} )}}{{{k_B}T}}$, it is easy then to see that the effective activation energy for dewetting ${E_a}$ is the slope of the straight fitting of Fig. 5(b).Fig. 5. Time evolution of the dewetted area fraction of 150 nm films treated at different isothermals (a). The dashed lines represent the data fitting by using the equation $X = 1 - exp( - [{R + K({t - \tau } ){]^2}} )$. Arrhenius plot $Ln({K{k_B}T} )$ Vs $1/{k_B}T$ $(b )$. Effective activation energy for dewetting of As20Se80 films with three different thicknesses (c).
The transformation curves of the film with 150 nm thickness grant us to obtain the effective activation energy for dewetting 157.57 ${\pm} \; $12.74 $KJ/mol$, which has the same order of magnitude as the values obtained by Yang et al. [22] for the activation energy for viscous flow and the surface diffusion in bulk glasses. The dewetting activation energy is also calculated for two other films with 240 and 513 nm thickness (Fig. 5(c)), which gives information on the transport mechanism by which dewetting occurs. Studies on relaxation dynamics and diffusion of molecular organic glasses have shown that the process of viscous flow and surface diffusion are decoupled phenomena, each one with different mechanisms of mass transport [23,24]. The structural viscous flow relaxation is influenced by the interactions with the substrate, while the surface diffusion coefficient is invariant with thickness. From this approach, since our results suggest thickness-dependent activation energy, we can say the process of dewetting in As20Se80 films is predominantly due to viscous flow rather than surface diffusion.
3.3 Heating rate influence
The evolution of the dewetted area fraction registered while heating the sample at different rates (Fig. 6) demonstrate the agglomeration process is also affected by the heating rate. As observed in the derivative of the experimental curves (bottom of Fig. 6) lower heating rates lead the dewetting of the film to occur at lower temperatures, and within ranges of temperatures slightly narrower. These aspects have a great impact on the final size and distribution of islands since higher heating rates cause the formation of droplets with wider size distribution and with higher mean diameters. As shown in Table 1, the heating rate also changes the symmetry of the distribution, and while the mean and the modes of the distributions are practically equal when heating at 10 and 50 °C/min, the use of higher rates produces less symmetric distributions.
Fig. 6. Dewetted area fraction as a function of temperature, while heating 225 nm thick films at different heating rates.
Table 1. Droplets size distribution and number of holes in films of 225 nm heated at different rates. The same size area is used as a reference to obtain the data.
The influence of the heating rate in the final material distribution comes from the nucleation process (holes appearance) and its evolution on time. The formation of holes is more uniform in time when heating at higher rates than when heating slowly. When heating at 100 and 150 °C/min, for example, the number of holes increases constantly in a wider range of temperatures causing the processes of nucleation and growth to occurs simultaneously. The optical images at the inset of Fig. 6, show that heating rapidly causes the simultaneous growth of holes of different sizes. This lack of homogeneity in the nucleation process causes an asynchrony in the growth of the holes that is reflected in variations of viscosity during the droplet formation. The droplets are formed at different viscosities and consequently are widely dispersed in size. On the other hand, heating slowly provokes most of the holes to appear at the same time (at lower temperatures) and to evolve synchronously with the same size along the time (see inset of Fig. 6). Slow heating rates induce the formation of less dispersed size distributions and smaller droplet sizes.
3.4 Atmosphere influence
As shown in Fig. 7 the interfacial interactions between the substrate and the atmosphere also have a great impact on the evolution of the dewetting. According to the atmosphere, the dewetting of 225 nm films showed droplets with different spatial distributions and sizes. Dewetting in a nitrogen atmosphere causes droplets to be arranged in a polygon pattern defined by heterogeneous nucleation (Fig. 7(a)); on the other hand, treatments carried out in oxygen atmosphere produce droplets uniformly dispersed on the substrate (Fig. 7(b)). Further analysis shows the droplets obtained in the oxygen atmosphere are spatially distributed in correlated positions according to the dewetting by capillarity instabilities or spinodal dewetting [25,26] (see the FFT of droplets distribution in the inset of Fig. 7(b)). Additionally, inducing dewetting in the oxygen atmosphere results in the formation of a higher number of droplets with smaller sizes in comparison with the dewetting in air or nitrogen atmosphere (Fig. 7(c)). The same tendency is observed no matter the heating rate or the temperature used for inducing dewetting.
Fig. 7. Pattern of droplets after dewetting of 250 nm films under nitrogen (a) and oxygen atmosphere (b). The inset of (b) shows the FFT of the position of the droplets. Number of droplets vs droplets diameter produced by dewetting in nitrogen, oxygen, and air atmospheres (c).
To unveil the origin of such variations it is important to analyze the surface interactions between the thin film and the respective atmospheres. For this purpose, the surface of films as-deposited and films treated with oxygen, and nitrogen atmospheres are analyzed by XPS (more details of the atmosphere treatment in the experimental section). Besides the arsenic and selenium elements, we also detected adsorbed oxygen. The analysis of the high-resolution spectra of Se 3d, As 3d, and O 1s core levels are summarized in Table 2.
Table 2. Binding energy (BE) of As 3d, Se 3d, and O 1s core levels and atomic composition of As20Se80 surfaces as-deposited, after thermal treatment under oxygen, and nitrogen atmospheres. The two components used to deconvolute the As 3d and Se 3d levels are assigned to different bonds inside of the glassy matrix, according to literature.
In all the samples, the width of the selenium 3d core level depicts the presence of selenium at different chemical bonds. As showed in Fig. 8, the Se level is deconvoluted using three doublets (spin-orbit splitting), each one with binding energy (BE) corresponding to the three possible bond configurations of Se in the glass structure. The main peaks (3d5/2) correspond with the values reported in the literature for selenium bonded to two arsenic atoms As-Se-As (peak at 54.83 eV), selenium bonded to both arsenic and selenium Se-Se-As (peak at 54.95 eV), and bonded to two other seleniums Se-Se-Se (peak around 55.39 eV) respectively [17,27]. These three peaks are invariant within the different samples, which indicates the selenium is unaffected by the atmosphere. This also agrees with energetically unfavorable oxidation of surface selenium [28].
Fig. 8. XPS high-resolution spectra of Se 3d (a) and As 3d (b) levels in as-deposited films and As 3d core level of films treated under oxygen atmosphere (c).
Differently than selenium, the arsenic and oxygen core levels differ greatly from sample to sample. In as-deposited samples, the high-resolution spectrum of As 3d level is deconvoluted using two spin-orbit doublets. Most of the arsenic is represented by the peak at around 43.0 eV (See Fig. 8(b)) that corresponds to arsenic bound to 3 selenium atoms Se-As< (Se)2 in the pyramidal structural units of the glass [17,27]. Furthermore, there is a second As component around 43.5 eV that shows a small fraction of the arsenic (1.5 at%) bound with slightly higher binding energies.
It has been demonstrated that under room conditions (light and atmosphere) arsenic-based surfaces endure photo-induced degradation followed by arsenic oxide formation [28–32]. However, the arsenic oxidation is not direct since the reaction starts at the arsenic dangling bonds present on the surface; at this stage, the arsenic atom remains bonded to the surface while simultaneously bonds with the oxygen in the atmosphere. Under extended exposition to room conditions, the arsenic eventually goes through complete oxidation and forms arsenic oxide As2O3 (BE ∼45 eV) [29,33,34]. Since our samples had little exposition to room conditions before the XPS experiment, we suggest the second As 3d peak (43.5 eV) corresponds to arsenic in a suboxide or arsenic-oxyselenide compound analogous to the presented by Kolobov et al. in AsS surfaces [34]. Here, the As atom is bonded to selenium while is simultaneously bonded to oxygen as follows: (O)x-As<(Se)3/2-x with 0<x<3/2. Due to the oxygen's higher electronegativity (compared with selenium), the binding energy of arsenic in this compound should lay between the one of As2Se3 units (43.0 eV) and the completely oxidized arsenic As2O3 (∼45 eV) [29,31,35]. The formation of an arsenic-oxyselenide is supported by the presence of a great amount of strongly bonded oxygen (See Table 2) and the As 3d component at 43.5 eV in as-deposited samples (Table 2 and Fig. 8(b)).
Samples treated under nitrogen are very similar to the as-deposited films; besides showing the presence of arsenic in the oxyselenide (O)x-As<(Se)3/2-x compound (BE = 43.4 eV), the variations of atomic composition are negligible. The result suggests that interactions between film and nitrogen atmosphere do not involve the formation of chemical bonds and thus, nitrogen can be considered as a neutral atmosphere for the dewetting of As20Se80 films.
The opposite is observed at samples treated under oxygen. While the Se 3d level remains unchanged, the As 3d component related to the oxyselenide disappears completely (Fig. 8(c) and Table 2. As mentioned before, the arsenic-oxyselenide in as-deposited films is just an intermediate state toward the complete oxidation of arsenic. We believe the thermal treatment under an oxygen atmosphere induced the complete oxidation of arsenic into more volatile arsenic oxides (As2O3 or As2O5). These oxides can not be detected in the XPS As 3d spectra due to their high volatility under the X-ray radiation and ultrahigh vacuum of the XPS measurements [29,36]. Although there is not oxyselenide detected in these films, the oxygen atomic percentage increases due to the physical adsorption of oxygen in the sample surface. It is important to highlight that the variations in the atomic concentration of elements as Se and As are not necessarily related to the loss or increase of material, but to the atoms present in the first 2 to 4 nm of the surface (XPS penetration depth). Being covered by a layer of adsorbed oxygen, the XPS signal of these elements can be attenuated.
The distribution of droplets after dewetting under oxygen (Fig. 7(b)) can be directly related to the oxidation of surface arsenic. The presence of oxygen promotes the formation of arsenic oxides and therefore the losses of arsenic due to evaporation of the volatile oxides. The spinodal dewetting that we propose to explain the correlated distribution of droplets can be trigger by the local variations in height due to arsenic evaporation.
3.5 Dewetting and photolithography
It is well reported that the main losses in chalcogenide waveguides and microstructures are due to the surface rugosity (specially side-wall rugosity) introduced by photolithographic processes [3]. Although the reflow of chalcogenide microstructures has been used to lower these losses in waveguides [37], we propose to use the dewetting of chalcogenide thin films in combination with standard photolithographic processes to assemble photonic structures with superior quality. This would not only lower the wall's rugosity and film stress but also reshape their cross-section [38,39]. Therefore, the integration of both techniques may lead to 3D profiles that are only achievable by gray-scale lithography. To demonstrate the effect of the dewetting, a set of microstructures were dry etched in a 750 nm As20S80 film and were submitted to a controlled dewetting according to the parameters stablished in the previous sections. Milder conditions, i.e. lower temperatures and longer times, are used to avoid capillarity instabilities which could cause the film rupture inside the longer or larger microstructures. In addition, the severity of the thermal treatment is dictated by the microstructure’s aspect ratio, as smaller structures tend to dewet at lower temperatures than larger ones given their relative high stress.
In the case of waveguide-like structures (Fig. 9(a)), the larger dimension is conserved intact whereas the cross-section suffers a drastic reshaping, going from rectangular to a parabola (Fig. 9(b)). It is also worth noting that, as the chalcogenide layer reflows, sharp edges and regions with high rugosity smooth out due to the viscous flow changing the apparent rugosity of the material after the dewetting. Before dewetting, the rugosity (Sq) measured on top of the waveguide is around 2.47 nm changing to 0.32 nm before the dewetting reshaping.
4. Conclusions
The dewetting of As20Se80 thin films with thicknesses between 150 and 750 nm is studied in terms of the film thickness, temperature, heating rate, and atmosphere of dewetting. Inducing the dewetting of films by using temperatures between 180 and 300 °C, produces droplets of material that, after cooling, have the same molecular structure and atomic composition as the original film. The size of the droplets can be coarsely tuned by changing the film thickness and can be finely adjusted by the temperature at which the dewetting is induced. As for the heating rate, it controls the hole nucleation process and thus the final distribution of droplet size.
The study of the kinetics of dewetting in different film thicknesses demonstrates that dewetting of As20Se80 thin films, in the range of temperatures used here, is not limited to the surface diffusion (as in solids), but it also occurs by the bulk mass transport due to structural relaxation of the glass matrix. On the other hand, the atmosphere-film interface also influences the final microstructures droplets of material, since the interaction with reactive atmosphere can lead to the formation of arsenic suboxides that later transform completely into volatile compounds.
The dewetting of As20Se80 chalcogenide thin films has been demonstrated to be a potential tool for the microfabrication of photonic structures. By understanding the kinetics and dynamics of the dewetting process, it is possible to control the capillarity instabilities and form high aspect ratio structures of great importance in integrated photonics.
Funding
Fonds de recherche du Québec – Nature et technologies; Canada Foundation for Innovation; Natural Sciences and Engineering Research Council of Canada.
Acknowledgments
This research was supported by the Natural Sciences and Engineering Research Council of Canada (NSERC), the Canada Foundation for Innovation (CFI), and the Fonds de recherche du Québec - Nature et technologies (FRQNT) and Photonics.
Disclosures
The authors declare no conflicts of interest
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
Supplemental document
See Supplement 1 for supporting content.
References
1. J.-L. Adam and X. Zhang, Chalcogenide Glasses Preparation, Properties and Applications (Woodhead Publishing Limited, 2014).
2. J. Sanghera and D. Gibson, “Optical properties of chalcogenide glasses and fibers,” Chalcogenide Glas. 2013, 113–138 (2013). [CrossRef]
3. B. J. Eggleton, B. Luther-Davies, and K. Richardson, “Chalcogenide photonics,” Nat. Photonics 5(3), 141–148 (2011). [CrossRef]
4. Z. Han, P. Lin, V. Singh, L. Kimerling, J. Hu, K. Richardson, A. Agarwal, and D. T. H. Tan, “On-chip mid-infrared gas detection using chalcogenide glass waveguide,” Appl. Phys. Lett. 108(14), 141106 (2016). [CrossRef]
5. P. Jean, A. Douaud, V. Michaud-Belleau, S. H. Messaddeq, J. Genest, S. LaRochelle, Y. Messaddeq, and W. Shi, “Etchless chalcogenide microresonators monolithically coupled to silicon photonic waveguides,” Opt. Lett. 45(10), 2830 (2020). [CrossRef]
6. S. Deckoff-jones, H. Lin, D. Kita, H. Zheng, D. Li, W. Zhang, and J. Hu, “Chalcogenide glass waveguide-integrated black phosphorus mid-infrared photodetectors,” J. Opt. 20(4), 044004 (2018). [CrossRef]
7. Y. Zhang, C. N. Woods, M. Alvarez, Y. Jin, R. A. Riggleman, and Z. Fakhraai, “Effect of substrate interactions on the glass transition and length-scale of correlated dynamics in ultra-thin molecular glass films,” J. Chem. Phys. 149(18), 184902 (2018). [CrossRef]
8. F. Leroy, Ł. Borowik, F. Cheynis, Y. Almadori, S. Curiotto, M. Trautmann, J. C. Barbé, and P. Müller, “How to control solid state dewetting: A short review,” Surf. Sci. Rep. 71(2), 391–409 (2016). [CrossRef]
9. J. C. Meredith, A. P. Smith, A. Karim, and E. J. Amis, “Combinatorial materials science for polymer thin-film dewetting,” Macromolecules 33(26), 9747–9756 (2000). [CrossRef]
10. S. Herminghaus, “Spinodal dewetting in liquid crystal and liquid metal films,” Science 282(5390), 916–919 (1998). [CrossRef]
11. D. Gentili, G. Foschi, F. Valle, M. Cavallini, and F. Biscarini, “Applications of dewetting in micro and nanotechnology,” Chem. Soc. Rev. 41(12), 4430–4443 (2012). [CrossRef]
12. T. Das Gupta, L. Martin-Monier, W. Yan, A. Le Bris, T. Nguyen-Dang, A. G. Page, K. T. Ho, F. Yesilköy, H. Altug, Y. Qu, and F. Sorin, “Self-assembly of nanostructured glass metasurfaces via templated fluid instabilities,” Nat. Nanotechnol. 14(4), 320–327 (2019). [CrossRef]
13. Y. N. Colmenares, S. H. Messaddeq, and Y. Messaddeq, “Studying the kinetics of microstructure formation through dewetting of As-Se thin films,” Phys. Rev. Mater. 5(1), 015605 (2021). [CrossRef]
14. G. Reiter, “Unstable thin polymer films: rupture and dewetting processes,” Langmuir 9(5), 1344–1351 (1993). [CrossRef]
15. G. Reiter, S. Al Akhrass, M. Hamieh, P. Damman, S. Gabriele, T. Vilmin, and E. Raphaël, “Dewetting as an investigative tool for studying properties of thin polymer films,” Eur. Phys. J. Spec. Top. 166(1), 165–172 (2009). [CrossRef]
16. M. S. Iovu, E. I. Kamitsos, C. P. E. Varsamis, P. Boolchand, and M. Popescu, “Raman Spectra of AsxSe100-x glasses doped with metals,” J. Optoelectron. Adv. Mater. 7, 1217–1221 (2005).
17. O. Kondrat, R. Holomb, V. Mitsa, M. Veres, and N. Tsud, “Structural Investigation of As-Se Chalcogenide thin films with different compositions: formation, characterization and peculiarities of volume and near surface nanolayers,” Funct. Mater. 24(4), 547–554 (2017). [CrossRef]
18. C. V. Thompson, “Solid-state dewetting of thin films,” Annu. Rev. Mater. Res. 42(1), 399–434 (2012). [CrossRef]
19. Y. Zhang, E. C. Glor, M. Li, T. Liu, K. Wahid, W. Zhang, R. A. Riggleman, and Z. Fakhraai, “Long-range correlated dynamics in ultra-thin molecular glass films,” J. Chem. Phys. 145(11), 114502 (2016). [CrossRef]
20. G. Reiter, “Dewetting of thin polymer films,” Phys. Rev. Lett. 68(1), 75–78 (1992). [CrossRef]
21. E. Jiran and C. V. Thompson, “Capillary instabilities in thin, continuous films,” Thin Solid Films 208(1), 23–28 (1992). [CrossRef]
22. G. Yang, O. Gulbiten, Y. Gueguen, B. Bureau, J.-C. Sangleboeuf, C. Roiland, E. A. King, and P. Lucas, “Fragile-strong behavior in the AsxSe1−x glass forming system in relation to structural dimensionality,” Phys. Rev. B 85(14), 144107 (2012). [CrossRef]
23. Y. Zhang and Z. Fakhraai, “Decoupling of surface diffusion and relaxation dynamics of molecular glasses,” Proc. Natl. Acad. Sci. U. S. A. 114(19), 4915–4919 (2017). [CrossRef]
24. Y. Zhang and Z. Fakhraai, “Invariant fast diffusion on the surfaces of ultrastable and aged molecular glasses,” Phys. Rev. Lett. 118(6), 066101 (2017). [CrossRef]
25. S. Yadavali, M. Khenner, and R. Kalyanaraman, “Pulsed laser dewetting of Au films: experiments and modeling of nanoscale behavior,” J. Mater. Res. 28(13), 1715–1723 (2013). [CrossRef]
26. R. Xie, A. Karim, J. F. Douglas, C. C. Han, and R. A. Weiss, “Spinodal dewetting of thin polymer films,” Phys. Rev. Lett. 81(6), 1251–1254 (1998). [CrossRef]
27. R. Golovchak, A. Kovalskiy, A. C. Miller, H. Jain, and O. Shpotyuk, “Structure of Se-rich As-Se glasses by high-resolution X-ray photoelectron spectroscopy,” Phys. Rev. B: Condens. Matter Mater. Phys. 76(12), 125208 (2007). [CrossRef]
28. J. A. Frantz, J. D. Myers, R. Y. Bekele, C. M. Spillmann, J. Kolacz, H. Gotjen, V. Q. Nguyen, C. C. McClain, and J. S. Sanghera, “Arsenic selenide thin film degradation and its mitigation,” Opt. Mater. Express 8(12), 3659 (2018). [CrossRef]
29. F. Rochet, C. Poncey, G. Dufour, H. Roulet, W. N. Rodrigues, M. Sauvage, J. C. Boulliard, F. Sirotti, and G. Panaccione, “The As-terminated Si(001) surface and its oxidation in molecular oxygen: an Si 2p and As 3D core-level study with synchrotron radiation,” Surf. Sci. 326(3), 229–242 (1995). [CrossRef]
30. B. J. Flinn and N. S. McIntyre, “Studies of the UV/Ozone oxidation of GaAs using angle-resolved X-ray photoelectron spectroscopy,” Surf. Interface Anal. 15(1), 19–26 (1990). [CrossRef]
31. S. Krishnaswami, H. Jain, and A. C. Miller, “Effect of oxygen on the photoinduced changes in the electronic structure of As50Se50 glass films,” J. Optoelectron. Adv. Mater. 3, 695–702 (2001).
32. J. S. Berkes, S. W. Ing, and W. J. Hillegas, “Photodecomposition of Amorphous As2Se3 and As2S3,” J. Appl. Phys. 42(12), 4908–4916 (1971). [CrossRef]
33. S. C. Ghosh, M. C. Biesinger, R. R. Lapierre, and P. Kruse, “X-ray photoelectron spectroscopic study of the formation of catalytic gold nanoparticles on ultraviolet-ozone oxidized GaAs (100) substrates,” J. Appl. Phys. 101(11), 114322 (2007). [CrossRef]
34. A. V. Kolobov, J. P. Badyal, and R. M. Lambert, “Novel photoinduced surface oxidation of an amorphous semiconductor: an XPS study of vitreous arsenic sulfide,” Surf. Sci. Lett. 222(2-3), L819–L824 (1989). [CrossRef]
35. J. T. Bloking, S. Krishnaswami, H. Jain, M. Vlcek, and R. P. Vinci, “Photoinduced changes in the surface morphology of As50Se50 chalcogenide glass films,” Opt. Mater. 17(4), 453–458 (2001). [CrossRef]
36. H. Viltres, O. F. Odio, L. Lartundo-Rojas, and E. Reguera, “Degradation study of arsenic oxides under XPS measurements,” Appl. Surf. Sci. 511, 145606 (2020). [CrossRef]
37. J. Hu, N. N. Feng, N. Carlie, L. Petit, A. Agarwal, K. Richardson, and L. Kimerling, “Optical loss reduction in high-index-contrast chalcogenide glass waveguides via thermal reflow,” Opt. Express 18(2), 1469–1478 (2010). [CrossRef]
38. P. Guo, C. Li, W. Huang, W. Zhang, P. Zhang, and T. Xu, “Thermal annealing of Ge-Se thin films and its influence on waveguide performance,” Opt. Mater. Express 10(1), 129 (2020). [CrossRef]
39. Y. Zhao, C. Li, P. Guo, W. Zhang, P. Xu, and P. Zhang, “Exploration of lift-off Ge–As–Se chalcogenide waveguides with thermal reflow process,” Opt. Mater. 92, 206–211 (2019). [CrossRef]
